pytorch/xla

Pull Requests Commits

Enable eager mode for PyTorch/XLA (#7234)

JackCaoG committed 1 year ago

6d48ff5b

Update jax to a stable version(0.4.29) for r2.4 (#7627)

bhavya01 committed 1 year ago

Verified 12e0d811

backport #7609: Run paged_attention test with megacore mode only on TPU v4 (#7628)

bhavya01 committed 1 year ago

Verified 85664631

Fetch submodules separately when torch pin is specified (#7616)

bhavya01 committed 1 year ago

Verified 7ed17578

Backport #7329: Avoid OOM when build PyTorch with CUDA enabled (#7618)

bhavya01 committed 1 year ago

Verified 844856d2

Add torch_pin to 2.4 release branch (#7336)

bhavya01 committed 1 year ago

Verified 0a17f095

[backport] Register custom call with CUDA plugin (#7303)

bhavya01 committed 1 year ago

Verified e44787a5

Update cuda_deps.yaml to include CUDA 12.4 (#7244)

bhavya01 committed 1 year ago

Verified 032e332e

[backport][Fori_loop|While_loop] Enable while_loop/fori_loop, add test case (#7157) (#7306)

ManfeiBai committed 1 year ago

Verified 27934620

[backport] Use np.prod instead of np.product (#7301) (#7310)

ManfeiBai committed 1 year ago

Verified a901eb8e

Revert the mul change (#7271) (#7285)

JackCaoG committed 1 year ago

Verified 72961330

Avoid log spam (#7278) (#7284)

JackCaoG committed 1 year ago

Verified fda5828e

Cherry-pick: PJRT plugin defaults (#7249 and #7268) (#7270)

will-cromar committed 1 year ago

Verified f00bd0b9

Backport: Bump jaxversion corresponding to libtpu 0612 (#7262)

vanbasten23 committed 1 year ago

Verified 5622efc4

Backport xla pin update (0612) (#7261)

vanbasten23 committed 1 year ago

Verified da5843c6

Use dest_offsets directly in LoadPlanner (#7243)

jonb377 committed 1 year ago

Verified 62425ebb

Add build trigger for 2.4.0-rc1 release

bhavya01 committed 1 year ago

7ee1bab6

[ZeRO-1] Sync features from r2.1_aws_neuron branch (#7132)

hgt312 committed 1 year ago

Verified 551a76cd

temporarily skip the pjrt test (#7230)

zpcore committed 1 year ago

Verified d0fb59e7

Use std::optional (#7228)

cyyever committed 1 year ago

Verified 2c71e9b3

torchbench CI (#7162)

zpcore committed 1 year ago

Verified 291764f6

Fix wrong env flag for using auto sharding in documentation (#7225)

lausannel committed 1 year ago

Verified 28c5e141

Remove stack dump for compiler failure which can happen often (#7207)

jeffhataws committed 1 year ago

Verified 5338a3b0

Copy examples dir for TPU CI (#7199)

JackCaoG committed 1 year ago

Verified 1ac8968d

Add stub plugins for XPU and Neuron (#7209)

will-cromar committed 1 year ago

Verified f9851b36

Remove unused variables from run_tests.sh (#7217)

bhavya01 committed 1 year ago

Verified de283c98

init glossary (#7214)

ahoblitz committed 1 year ago

Verified 66c7ee8b

Training example for llama3; misc changes to make the training work. (#7194)

qihqi committed 1 year ago

Verified f4a612cb

Add int8 per channel weight-only quantized matmul (#7201)

lsy323 committed 1 year ago

Verified 56ddd5de

bhavya01 committed 1 year ago

Verified 4a30ea7d

Older